How Tracking Workers With AI Could Raise Privacy Concerns

Employers are increasingly tracking their workers using artificial intelligence, and observers say the practice raises privacy concerns. 

Amazon installed machine learning-powered cameras in its delivery vans earlier this year. The company recently told its drivers they must agree to use them. The practice of on-the-job surveillance may be legal, but not everyone agrees it's ethical. 

"Companies have used employee monitoring tools for decades, but with advancing technology, they are becoming more invasive," Aimee O’Driscoll, a security researcher at technology comparison website Comparitech, said in an email interview.

"Most employee monitoring could be considered an invasion of privacy, and it could be argued that what Amazon is doing isn’t a whole lot different from having CCTV cameras at an office." 

Be Watched or Get Fired

Amazon’s approximately 75,000 delivery drivers in the US are now required to sign a "biometric consent" form. The permission form lets AI-powered cameras watch drivers' location, movement, and biometric data. Employees who don’t sign the document could be fired. 

AI surveillance tracks everything the Amazon drivers do, invading the subject's assumed privacy, Chris Hauk, a privacy researcher at the privacy website Pixel Privacy, said in an email interview. 

"This includes recording every time the subject yawns or scratches a sensitive area," he added. "A driver has at least some right to a semblance of privacy while inside their vehicles."

Employee surveillance by AI is a growing issue. Walmart has patented an AI technology that enables listening to employee interactions with customers at checkouts, O’Driscoll pointed out. 

Software developer Enaible provides productivity tools based on AI software. Various companies, including Macy’s, use Microsoft’s analytics system, which can monitor employee behavior. Domino’s has used AI technology to check that pizzas are made correctly.

"One of the main issues is that when used in the wrong way, lack of privacy can lead to security issues," O’Driscoll said.

"Surveillance data could end up in the hands of criminals, or bad actors themselves could use AI surveillance to target victims."

Biometric face scans result in a mathematical representation of the data subject’s face that can be leveraged to identify and track them anywhere, for the rest of their life, Ray Walsh, a privacy expert at the website ProPrivacy, said in an email interview.

"This raises serious privacy and security risks for Amazon drivers whose data could be breached, leaked, or potentially even leveraged by government snoops using a warrant," he added. 

Because AI technology is relatively new, there aren’t concrete guidelines around its use concerning privacy, O’Driscoll said. "That said, safety will likely always be a legitimate excuse for privacy invasion, so Amazon is covered under its current intentions," she added. 

O’Driscoll said that AI surveillance should be regulated. "Companies should ideally have legitimate reasons (such as safety) for AI surveillance," she added. 

Some states have passed bills that limit and regulate the use of surveillance in the workplace, but federal legislation is lacking, Walsh pointed out.

In 2019, the Algorithmic Accountability Act was introduced into the House and Senate, but it was ultimately rejected. A similar act may be introduced this year, and "it is hoped that a Democrat-led Congress and White House may improve its chances of passing," Walsh said. 

Don’t Expect Privacy on the Job

Not everyone thinks that Amazon’s surveillance of its drivers is an invasion of privacy. 

"I think it is hard to argue that an employee has a reasonable expectation of privacy in a company-owned vehicle," Will Griffin, the chief ethics officer for AI company Hypergiant, said in an email interview.

"The bigger concern is that in a few short years, all of these drivers will be replaced by autonomous vehicles. So any debate about driver policy will become a moot point as the fleet becomes fully autonomous."

The Amazon case highlights the need for a union or targeted federal intervention as a way to restore the balance of power between Amazon's workers and the company, Griffin said.

"While employees have the right to quit or consent to the proposed AI monitoring, they are doing so in the face of an enormous power imbalance," he added.